Childhood Trauma, Brain Structure and Emotion Recognition in Patients With Schizophrenia and Healthy Participants

Karolina I. Rokita; Laurena Holleran; Maria R. Dauvermann; David Mothersill; Jessica Holland; Laura Costello; Ruán Kane; Declan McKernan; Derek W. Morris; John P. Kelly; Aiden Corvin; Brian Hallahan; Colm McDonald; Gary Donohoe

Disclosures

Soc Cogn Affect Neurosci. 2020;15(12):1336-1350. 

In This Article

Results

Demographic and Clinical Data

Sociodemographic and clinical characteristics of the study participants are presented in Table 1. Patients did not significantly differ from healthy participants on age (P = 0.063) and gender (P = 0.838). As expected, both groups showed significant differences on years of education (P < 0.001) and estimated IQ (P < 0.001).

Between-group Differences on Childhood Trauma, Emotion Recognition and ROI Volumes

Patients with SZ had significantly higher scores on physical neglect (P = 0.018), but not on emotional abuse (P = 0.187), physical abuse (P = 0.463), sexual abuse (P = 0.069) or emotional neglect (P = 0.560) compared to healthy participants (Table 1). Both groups also significantly differed on the total childhood trauma score (P = 0.049). As expected, patients showed significantly poorer performance than healthy participants on the ERT task (P < 0.001; Table 1).

When comparing patients and healthy participants on ROI volumes, we found that only the hippocampal volumes were significantly smaller in patients compared to healthy participants, including the total (P = 0.001) as well as the left and right hippocampal (P = 0.003; P = 0.001, respectively; Table 2) volumes. No significant differences on the amygdala and ACC volumes between patients and healthy participants were observed.

Associations Between Childhood Trauma, Emotion Recognition and ROI Volumes

All Participants. To assess the relationship between childhood trauma, emotion recognition and ROI volumes and to select the variables for the mediation analyses, partial correlations of these variables are reported for all participants (Table 3).

Regarding the associations between 'childhood trauma and ROI volumes', CTQ total score was significantly negatively correlated with the total ACC volume (r = −0.161, P = 0.048) (Figure 1). There was also a trend towards significance for the association between physical neglect and the total ACC volume (r = −0.158, P = 0.052) (Figure 2). Total CTQ scores were also negatively correlated with hippocampal and amygdala volumes, but these correlations were non-significant.

Figure 1.

Association between the total ACC volume and CTQ total score (log transformed) in all participants. The scatter plot is unadjusted for covariates for illustration purposes (r = −.184, p = .022).

Figure 2.

Association between the total ACC volume and physical neglect (log transformed) in all participants. The scatter plot is unadjusted for covariates for illustration purposes (r = −.204, p = .011).

In terms of the associations between 'performance on the ERT task and the ROI volumes', the ERT score was significantly positively associated with the total (r = 0.262, P = 0.001), left (r = 0.206, P = 0.011) and right (r = 0.224, P = 0.005) ACC volumes as well as the total (r = 0.172, P = 0.032), left (r = 0.168, P = 0.037) and right (r = 0.162, P = 0.044) hippocampal volumes. No significant correlations were reported between the amygdala and ERT task. The ERT score was also significantly negatively associated with physical neglect (r = −0.321, P < 0.001), but not total childhood trauma score (r = −0.134, P = 0.098).

Healthy Participants. In subsequent correlational analyses conducted in healthy participants only, physical neglect and CTQ total score were significantly negatively correlated with the left ACC volumes (r = −0.203, P = 0.037; r = −0.206, P = 0.034, respectively). Total CTQ score was also significantly negatively associated with the total ACC volumes (r = −0.227, P = 0.019). No significant associations were reported with hippocampal and amygdala volumes.

In terms of the association between performance on the ERT task and the ROI volumes, the ERT score was significantly positively correlated with the total (r = 0.313, P = 0.001), left (r = 0.233, P = 0.016) and right ACC volumes (r = 0.281, P = 0.003). No significant correlations were reported with hippocampal and amygdala volumes. The ERT score was also significantly negatively associated with physical neglect (r = −0.215, P = 0.025), but not total childhood trauma score (r = −0.175, P = 0.069).

Patients With SZ. When conducting correlational analyses in patients only, physical neglect and CTQ total score were not significantly associated with any of the ROI volumes. However, a significant positive association was found between the ERT task and left hippocampal volume (r = 0.325, P = 0.034). The ERT score was also significantly negatively associated with physical neglect (r = −0.317, P = 0.038), but not total childhood trauma score (r = 0.131, P = 0.403).

Additionally, we examined the possible impact of antipsychotic medication on brain structure and behavioural performance in patients with SZ, however, we found no significant correlations, indicating that findings were independent of medication in our sample.

Moderated Mediation Analyses

In order to test our main hypothesis regarding the direct and indirect effects of physical neglect and the CTQ total score on emotion recognition as mediated by volumetric changes, we tested two moderated mediation models in which group (healthy participants and patients separately) represented a moderator variable, and age, years of education and ICV were included as covariates. Based on the correlational analyses, only the ACC regions were entered as mediators in both models as they were associated with childhood trauma as well as emotion recognition measures. Physical neglect and the CTQ total score served as independent predictor variables and total ERT score was selected as the outcome (dependent) variable (Table 4). Additional analyses with the amygdala and hippocampal volumes as mediators were subsequently conducted to ascertain the possible mediating effects of these regions. A significant effect of diagnosis as a moderating variable was observed (F = 5.6; df1,148; P = 0.019); therefore, the mediating effects of brain volume are reported separately for healthy controls and patients.

Healthy Participants. In healthy participants, significant indirect effects were observed in the association between the CTQ total score and the ERT task scores by total, left and right ACC volumes (Total ACC: b = −0.0206, SE = 0.0101, 95% CI [−0.0437: −0.0045]; Left ACC: b = −0.0145, SE = 0.0087, 95% CI [−0.0353: −0.0012] and Right ACC: b = −0.0119, SE = 0.0092, 95% CI [−0.0356: −0.0001], respectively; Table 4 and Figure 3). We also found evidence of full mediation of the association between physical neglect and ERT total scores by both total and left ACC volumes (Total ACC: b = −0.0692, SE = 0.0431, 95% CI [−0.1779: −0.0083] and Left ACC: b = −0.0643, SE = 0.0425, 95% CI [−0.1699: −0.0034]; Table 4 and Figure 4). In additional analyses with the amygdala and hippocampal volumes as mediators, no significant indirect effects were observed in healthy participants.

Figure 3.

A simplified moderation mediation model on the mediating effects of the total and left ACC in the relationship between the CTQ and ERT total scores, controlling for age, years of education and ICV.
Note: Solid lines represent healthy controls (HC) and dashed lines represent patients. All presented effects are unstandardized; b = unstandardized coefficient; CTQ: Childhood Trauma Questionnaire; ACC: anterior cingulate cortex; ERT: Emotion Recognition Task.

Figure 4.

A simplified moderation mediation model on the mediating effects of the total and left ACC in the relationship between physical neglect and the ERT total score, controlling for age, years of education and ICV.
Note: Solid lines represent healthy controls (HC) and dashed lines represent patients. All presented effects are unstandardized; b = unstandardized coefficient; ACC: anterior cingulate cortex; ERT: Emotion Recognition Task.

Following these mediating effects, the direct pathway of CTQ total score (b = −0.0207, SE = 0.0346, P = 0.5512, 95% CI [−0.0890: 0.0477]) and physical neglect score (b = −0.1803, SE = 0.1633, P = 0.2715, 95% CI [−0.5030: 0.1425]) on emotion recognition were not significant (indicating full mediation). After inclusion of gender as another covariate in addition to age, years of education and ICV, all results remained significant except for the mediating effects of the right ACC volume in the association between the CTQ total score and the ERT task.

Patients With SZ. In patients, a significant direct association was found between physical neglect and the ERT task (b = −0.5597, SE = 0.1856, P = 0.003, 95% CI [−0.9265: −0.1928]), but not between the CTQ and ERT total scores (b = 0.0082, SE = 0.0391, P = 0.8337, 95% CI [−0.0691: 0.0855]) (Table 4, Figures 3 and 4). No indirect (mediating) effects were observed in this group for the ACC volumes. In additional analyses with the amygdala and hippocampal volumes as mediators, no significant indirect effects were observed in patients. After the inclusion of gender as another covariate in addition to age, years of education and ICV, direct effects of physical neglect on the ERT task remained significant.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....